Automated Epileptic Seizure Detection by Using MLP-GA Model

Abstract One of the most often utilized signals for assessing the electrical activities of the brain is the Electroencephalograph (EEG) signal. The manual approach of determining epileptic irregularities is time-consuming, and the outcome may vary depending on the reader's level of expertise. In order to identify seizures from EEG signals, this research introduces a unique computer-aided approach that uses hybrid machine learning techniques. Discrete Wavelet Transform is used to extract features from the signals (DWT). A classifier is utilized to distinguish between the normal and seizure classes using a mix of the Self Organizing Neural Network (SONN) and Multilayer Perceptron (MLP) method. The performance of the suggested strategy produced a 99.8 % accuracy.